Re: draft-yergeau-rfc2279bis-02.txt for STANDARD

["Steven M. Bellovin" <smb@research.att.com>]

In message <202CFE2B-24B3-11D7-8818-0003934B2128@cisco.com>, =?ISO-8859-1?Q?Pat
rik_F=E4ltstr=F6m?= writes:
>
>On fredag, jan 10, 2003, at 13:03 Europe/Stockholm, Steven M. Bellovin 
>wrote:
>
>> From what others have said -- I know little of UTF
>
>Short description (just in case):
>
>When you encode a character from ISO-10646 or Unicode into UTF-8, the 
>length of the output depends on what value the codepoint has. It is a 
>variable length encoding.
>
>The wording I want should say the encoding is not limited to what is 
>described in the document, but can be longer, so people should not do 
>guesses on the size of the output (without inspecting the input, one 
>can calculate the size needed in one run, and then encode on the 
>second) just because the document has such a limit.
>

Right, that was what I assumed, and why I worded my suggestion the way 
I did -- that the encoding could produce longer output than 4 bytes, 
and that this is a danger.

		--Steve Bellovin, http://www.research.att.com/~smb (me)
		http://www.wilyhacker.com (2nd edition of "Firewalls" book)